Is there ways to set the covariance matrix as diagonal?

[CRquantum]

Great work!
Sometime when do GMM clustering, we assume the covariance matrix is diagonal. Just curious, is there ways in the code so that we can limit the covariance matrix as diagonal matrix?
Thanks much!

Best regards,
Rong

[Jonas-Finkler]

Hi Rong,
by now this functionality is not implemented.
The simplest way to implement it, would be to just calculate a diagonal covariance matrix in the M subroutine by considering each dimension individually.
There might be performance gains for high dimensional data if one would only store the diagonal entries of the covariance matrix instead of the full matrix, but this would require modifications in other places of the code.

Please let me know, if you need such a feature and I can help with the implementation.

Best wishes,
Jonas

[CRquantum]

Thank you so much Jonas!
Yeah, not urgent, when convenient for you, if you may be able to add this diagonal covariance matrix functionality it would be great, I think it could be very useful.
Best regards,
Rong